Twitter introduced a new policy for dealing with disinformation during times of crisis on Thursday, creating new guidelines for gating or restricting the promotion of certain tweets if they are seen to be propagating misinformation.
In a blog post announcing the new policy, Twitter’s head of safety and integrity, Yoel Roth, explained that “content moderation is more than just leaving up or taking down content,” and that “we’ve expanded the range of actions we may take to ensure they’re proportionate to the severity of the potential harm.”
False reporting of events, false charges involving weapons or use of force, or larger misrepresentation about atrocities or international response are all targets of the new strategy.
During emergencies, hoax tweets and other disinformation frequently go viral as individuals rush to spread unconfirmed information. The sheer rapidity of events makes typical verification or fact-checking techniques impossible to apply, posing a substantial problem for moderators.
Misinformation-related tweets will not necessarily be removed or banned under the new policy; instead, Twitter will include a warning label asking users to click a button before the tweet can be viewed (similar to the existing labels for explicit imagery). Algorithmic promotion of the tweets will also be disabled.
The stricter criteria are only intended for special occasions. The policy will be applied to information related to Russia’s current invasion of Ukraine at first, but Twitter hopes to extend it to all emergent situations in the future. Crisis is defined as “situations in which there is a widespread threat to life, physical safety, health, or basic subsistence” for the purposes of the policy.
The policy is being implemented at a crucial moment for Twitter, while the company’s authorized sale to Elon Musk remains in doubt. Musk has promised to reduce the company’s moderation mechanisms in favor of a maximalist vision of free expression. It’s unclear when or how Musk’s plans will be executed, since he claims the deal is on hold pending a bot probe.